Profile: Michael Becich, Rob Hartman, Joe Marks & Donald Taylor, The Pittsburgh Health Data Alliance

the four interviewees

Michael Becich
Chairman and Distinguished University Professor, Department of Biomedical Informatics
Professor of Pathology, Bioengineering, Computing and Information and Clinical/Translational Sciences
Associate Vice Chancellor for Informatics in the Health Sciences
Director, Center for Commercial Application of Healthcare Data
Associate Director, University of Pittsburgh Cancer Institute
Associate Director, Clinical and Translational Science Institute
University of Pittsburgh

Rob Hartman
Senior Manager of Business Development
UPMC Enterprises

Joe Marks
Executive Director, Center for Machine Learning and Health
Carnegie Mellon University

Donald Taylor
Assistant Vice Chancellor, Health Sciences Translation
Associate Professor of Biomedical Informatics, School of Medicine
Executive Director, sciVelo, Innovation Institute
Codirector, Center for Commercial Applications of Healthcare Data
Codirector, Clinical and Translational Science Institute, Innovation Core
Associate Director, Center for Medical Innovation
Associate Professor, Bioengineering, Swanson School of Engineering
Associate Professor, Plastic Surgery, School of Medicine
Faculty, McGowan Institute for Regenerative Medicine
University of Pittsburgh


[Editor: The Pittsburgh Health Data Alliance (PHDA) is comprised of three entities: UPMC Enterprises (UPMC’s innovation arm), CMU’s Center for Machine Learning and Health (CMLH), and the University of Pittsburgh’s Center for the Commercial Application of Healthcare Data (CCA).]


The Pittsburgh Health Data Alliance began March 2015. It was announced by the leaders of UPMC, CMU and Pitt as a way to unleash the potential of big data in healthcare. It was an agreement that included UPMC, CMU and Pitt to link the complementary strengths of the institutions: CMU’s strengths in computer science and artificial intelligence; Pitt’s biomedical informatics research and translational science research capabilities; and UPMC providing the clinical system and bringing the innovative payer-provider laboratory it has built to the table.

UPMC wants to be attached to and provide the next generation kind of care and this is one way that we can get involved early in technologies that will benefit our doctors, our providers every level and our patients. Saving lives and really improving the quality of care, reducing the cost of care, improving patient satisfaction, improving patient safety, … all of these are macro-goals of the PHDA.


The goal of the Health Data Alliance is to solidify Southwestern PA’s reputations as a hub of innovation in general and as a premiere hub of innovation in healthcare technology in particular. We want to support and retain academic talent and great students within the universities, to spin out startup companies that can live in the region and grow in the region, and to attract industry support to the region for healthcare technology.

What we would like to be able to tell entrepreneurs is “PHDA is the premier source of funding, support, expertise, data, follow-on funding, general hand-holding for digital health innovation.” and then to live up to that promise over time and get people to understand, not from what we say, but from what we’ve done.


The idea is to close the loop on innovation. Typically, researchers think about things, then write grant proposals. Those grants turn into experiments, which result in intellectual property, then we write papers. But the loop hasn’t been closed until you take that technology, embed it in a commercial setting, either by licensing to an industry partner or forming a company. Then the technology becomes real. Instead of serving a couple laboratories and maybe a hundred people, it gets out to thousands, millions, or more in terms of impact.


We would love to see commercial successes. Startups in the region is our primary endpoint for success and so if we could have 5-10 that would be a homerun. That’s an aggressive goal, but even if we had something on the order of 2-4 for startups per year, that would be a great thing for the region.

We see a wide range of successful outcomes. It can be Enterprises mentoring the company to help find external investment and/or putting in resources like data, to Enterprises itself taking an equity position in the company, all the way up to Enterprises taking a significant portion of the company in terms of capitalizing it and getting it off the ground.



We have open calls for proposals which we advertise to the faculty across all of the different schools and departments and institutes at CMU. We review proposals both for scientific rigor – and to ensure that they’re advancing the state-of-the-art. UPMC also conducts a review process and we agree on funding proposals by consensus. We also have we call “challenge problems:” real-world problems in the healthcare space that UPMC confronts every day in both its provider and insurance businesses. In that case, we hold events or even one-on-one meetings to make the faculty aware and then get them excited. So, it’s a mix of open calls that tap into the wild creativity of the faculty, along with alerts to more targeted, immediate problems that also require research solutions.

A project that we affectionately called “Tummy Tunes” was proposed by a psychologist, an artist and a statistician — sounds like the beginning of a joke, but actually an awesome team. They want to develop technology for monitoring tummy gurgles and see how that relates to mental state. They actually want to start a new field of psycho-gastroenterology. We talked to some colorectal surgeons about how they analyze gut sounds now. They think gut sounds contain a lot of information, so they met with the scientists and came up with at least three different use cases. In at least two of them, there’s clear financial impact for the healthcare system and clear impacts on the patient experience in the hospital or just the patient managing their own care.

Another exciting one is the Tumor Driver Identification (TDI) project which is focused on precision therapy for patients with cancer. It is evolving to a startup called Dionex and it’s taking a kind of three-pronged approach. One prong looks at the genomic variations that the tumors have. The second is looking at the immune system in patients and finding out which parts are turned on or turned off – having certain parts of the immune system turned on can be a good thing for the patient and a bad thing for tumors. The third looks definitively the systems of biology that inform what is the metabolic state of the tumor. Those things in combination can then customize therapy for patients with melanoma.


We’re looking to de-risk science in the context of something that’s durable and sustainable in the marketplace. So, the researchers have to factor in market dynamics, competitive landscape, intellectual property considerations, even regulatory and reimbursement issues, and bake all of these factors in at the earlier stages in the research project so that they are designing something that not only has scientific merit, but is useful and can hopefully more rapidly evolve into new companies.

The CCA has a partnership the Innovation Institute, the epicenter for innovation, entrepreneurship and commercialization at the University of Pittsburgh. When early potential projects are identified, one of the first things we do is file an invention disclosure that will characterize and really circumscribe what is the potential invention, what is the market landscape, how it has been funded so far, etc. Then we begin to do some background and intellectual property diligence work.


A major collaborator of the CCA is Pitt’s Clinical and Translational Science Institute or CTSI. Neptune is a research data warehouse that we built to support electronic medical record research data marts, in part for the PHDA. It is running through a service called health Record Research Request or R3. It supports investigations ranging from health care operations to patient safety to advanced, almost esoteric approaches to causal discovery modeling for identifying what targeted therapies might be most appropriate for individualized patient delivery. Data requests go through stage gates to make sure the right people are getting the right data and everything is aligned. R3 is currently assisting UPMC Enterprises with provisioning electron health records (EHR) to PHDA funded projects.

There’s a number of challenges: identifying all the data that is needed to support a PHDA project, then integrating, curating, sufficiently de-identifying, and then sharing the data. But it’s doable and we have done it successfully.


The School of Computer Science at CMU, the Department of Biomedical Informatics, the Department of Computational and Systems Biology, and the new School of Computing and Information at Pitt are producing tremendous talent. Our trainees find jobs immediately and generally get to choose from a variety of jobs. There’s a little bit of downside, however: our senior talent gets recruited to leadership positions at other universities, academic health systems, health care partners and innovative companies on a regular basis.

Up Next: Profile: Susie Puskar, Partner4Work

View all Interviews